It has been known in the art of superconductors for some years that intermetallic compounds of the A-15 type, perhaps most notably Nb.sub.3 Sn and V.sub.3 Ga, have superconductive properties which are very useful in the production of high magnetic fields, in superconductive motors and generators, and in other devices of this type. However, these compounds are metastable which fact prevents them from being made by ordinary chemical processes, and, once made, they are extremely brittle, which does not permit them to be handled or mechanically formed to any great extent. Furthermore, it is well known in the art that superconductors of a given type have more useful properties when they are disposed in the form of continuous filaments extending from one end of the superconductor to another; preferably the filaments are on the order of 2 microns in diameter. Since it is well-nigh impossible to form such filaments individually, it has become the practice of the art to form such fine filaments by reducing a matrix containing a number of rods of the desired compound or of a precursor of the desired compound from a comparatively large billet down to a very fine wire size; the rods are proportionately reduced. It is possible to further reduce the size of the filaments and increase their number by stopping the reduction process at an intermediate stage, cutting the worked composite into lengths, stacking these lengths in an extrusion billet, extruding this to a long rod and further drawing this rod down to wire. This process is used in the manufacture of superconductors of several types, both those of the A-15 type and those of the alloy type, such as NbTi superconductors, which are usually made in a Cu matrix. The A-15 conductors have been most profitably made in the past by employment of the so-called bronze process in which rods of Nb, for example, are embedded in a matrix consisting of Cu and up to about 14% Sn. This composite can be worked, although annealing is necessary from time to time as the bronze work hardens, to a fine wire size. At a desired final size, the composite is heat treated, typically for 50 hours at 700.degree.C., at which time the Sn from the bronze diffuses onto the interfaces between the Nb rods and the matrix and Nb.sub.3 Sn is formed. This process is disclosed, for example, in U.S. Pat. No. 3,728,165.
It is well known in the art that, despite considerable improvements in manufacturing and in scientific understanding of the principles of superconductors, from time to time the superconductive property will break down. Ordinarily, this is due to an exceeding of any one of several critical parameters which are different for each superconductor: the critical temperature T.sub.C, the critical magnetic field H.sub.C and the critical current J.sub.C . Most commonly, perhaps, the critical field is exceeded causing the superconductive property in some of the filaments to be lost; the filaments are said to "go normal". This causes the current, of course, in the other filaments to be increased which may result in an exceeding of their critical current with possibly catastrophic results. Therefore, it is well known that a quantity of a suitable electrical conductor must be provided in parallel with the superconductor so that should some of the superconductor filaments "go noraml" an alternative current path or shunt is provided, thus providing a time during which the superconductive property can be restored. Since this phenomenon of going normal is ordinarily local, it is not sufficient to provide a shunt of a good electrical conductor external to the wire as then the heat generated in the normal conductor would be very great. Rather, it is better to provide the normally conductive material in close proximity to the wire to the superconductive wire and, in fact, it is generally accepted that the best way is to provide the wire itself with a quantity of highly electrically conductive material as part of its structure. Thus, wires have been made, typically with a central core of Nb rods having Nb.sub.3 Sn on their surfaces all embedded in a bronze matrix, and having an external layer of stabilizing high conductivity copper on their outer surface. See for example U.S. Pat. No. 3,728,165. It is desirable that the copper be on the outer surface of the wire rather than within it because the superconductive wires, operating as they do in a very low temperature environment, are usually bathed in liquid helium, so as to bring the temperature of the wire down to a point where the superconductive property is observed. Since the superconductor itself has no electrical resistance, under normal conditions no heat will be generated by the wire. However, if any fraction of the superconductive filaments go normal and current must therefore be carried in the normally conductive material, heat will, of course, be generated due to Joule heating. If the normally conductive material is on the outside of the conductor and this is in a bath of liquid helium at approximately 4.2 K, any heat generated in the normally conductive material can readily be carried off by the liquid helium.
One difficulty which has been faced and partially solved in the prior art is that during heat treatment the Sn or Ga from the bronze matrix tends to diffuse not only onto Nb to form Nb.sub.3 Sn but outwards into the Cu in order to alloy the Cu into a bronze. This effect, even when of minor quantitative extent, reduces the electrical conductivity of the Cu to such a degree that it is not suitable for its intended purpose. Therefore, some means must be found to protect this Cu from the diffusion of Sn or Ga. The most common method used in the prior art has been to interpose a layer of Ta barrier material, which is impermeable to Sn, between the matrix and the Cu stabilizer. See British Pat. No. 1,394,724. Other methods have been tried: for example layers of Nb have been used; refinements of this method include adding other elements to the Nb so that Nb.sub.3 Sn does not form on this layer thus making it too a superconductor, as disclosed, for example in British Pat. No. 1,499,507. However, these methods are troublesome and in most cases complicated and expensive.
While this method of interposing a Ta barrier between the bronze matrix and the Sn stabilizer is of great utility and has been generally accepted by the manufacturers of superconductors, it is not an easy or inexpensive method to practise due to various qualities of the Ta itself, so that the art has felt a need for improvement in this method.
The difficulties caused by the use of Ta as a diffusion barrier are several. For example, Ta is not a common material and is not readily available in the forms that are desired by the manufacturer of superconductors. It would be best from an economic standpoint to make a stabilized superconductor by first providing an extrusion billet of bronze having Nb rods disposed therein, and extruding this to a long, comparatively thin rod, preparatory to drawing the conductor down to its final size. Ideally, the tantalum barrier would be applied in tube form at some intermediate size and this assembly inserted into a Cu tube, so that the Ta would not have to be extruded. However, Ta is not commonly available as tubing, due to the fact that it is very refractory and hard to work; this results in a limited selection of standard sizes of tubing being available. It has turned out in practice that the Ta barrier layer ordinarily must be added to the matrix before extrusion. For purposes of illustration it can be noted here that the typical extrusion billet for a large quantity of superconductor will be on the order of 200 mm in diameter and 1 m long. After extrusion, such a billet might be 75 mm in diameter and 7.1 m long. Clearly if the stabilizing Cu and Ta could be added after extrusion a great deal more superconductive material could be extruded at once, which would be much more efficient. However, Ta not being available as tubes suitable to fit over an as-extruded rod means that tubing must be manufactured to the size of the billet. This is because tubing can be manufactured in nonstandard sizes only by spinning it, using conventional spinning processes, from the flat. However, spinning is limited in that a long tube cannot be formed since a machine tool must penetrate and work on the inside of the tube as well as on the outside. Clearly, a tube 7 m long by 75 mm in diameter simply cannot be made by spinning; however, one 1 m by 200 mm in diameter can be, and it is for this reason that the Ta barrier layer has been in the past generally applied prior to extrusion.
The suggestion has also been made that it would be possible to make Ta tube of any given length and diameter by simply forming a strip of suitable width and length into a tube and welding up the seam. However, as is frequently the case with pure metals, the welding process tends to result in a very large grain size in the area of the welded joint. Tantalum being hard and refractory, these grains will eventually tend to deform unevenly under working; that is, they will tend to deform along the preferential slip planes of the various large grains, resulting in a crack at or near the weld. This means that when the final conductor is heat treated to form Nb.sub.3 Sn from the Nb rods in the bronze, the Sn will diffuse through the crack in the Ta tube and diffuse into the Cu, thus lowering its electrical conductivity and rendering it more or less useless as a stabilizing material. Hence, the welding method has not been employed.
A further difficulty with the use of Ta as a barrier layer is that it is, as mentioned above, a very hard and refractory material, and as it is worked, it work hardens to a considerable extent, as does the bronze of the matrix. However, while the bronze of the matrix can be annealed relatively easily and at a comparatively low temperature, tantalum can only be annealed at such high temperatures that the bronze of the matrix would actually melt. Hence the tantalum as it is worked simply gets harder and harder and eventually inevitably cracks, again allowing the tin to diffuse into the stabilizing copper and destroying its effectiveness as a stabilizing material. Clearly if the Ta can be added after extrusion, that fraction of the work-hardening is avoided.
A final difficulty with Ta is that it does not bond metallurgically to itself. This prevents a homogenous layer from being formed by simply wrapping a Ta sheet around the extruded rod; this in turn results in uneven stress concentrations in the area of any overlap during drawing; cracking is again the eventual result.